Addressing Uncertainty in Flood Hazard Mapping under a Bayesian Approach

Abstract

Flood mapping is a crucial tool for assisting urban planning and emergency response plans and, consequently, preventing or reducing the risks associated with flood disasters. However, in developing countries that often lack or have limited data, to produce such maps is a challenging task. When topographic data are lacking, digital elevation models (DEMs) derived from the Shuttle Radar Topography Mission (SRTM) are frequently used as a freely available surrogate, albeit with additional uncertainty. This work presents an integrated framework to investigate flood inundation areas using a Bayesian approach, while including steps for calibrating SRTM data and determining the river bathymetry below the WSE. A flood event in the Itaqui municipality, in the state of Rio Grande do Sul, southern Brazil is used to demonstrate the proposed framework. Findings suggest benefits in using calibrated SRTM DEMs for flood mapping regardless of whether flood inundation areas were derived directly from projections of WSEs on the terrain or based on hydraulic simulations. Results further highlight the potential of using a Bayesian approach to improve quality and reliability of flood hazards maps, especially in regions that lack topographic data.